Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

GPFX is one of the new NQs, and it exhibits greater hepatobiliary transport than other NQs (Akiyama et al., 1995). As a first step in clarifying the mechanism that determines the degree of biliary clearance of NQs, we studied their hepatic uptake by isolated rat hepatocytes and found that NQs are taken up by the hepatocytes via a carrier-mediated active transport system. GPFX uptake is much greater than that of other NQs (Sasabe et al., 1997). Furthermore, the biliary excretion of GPFX was reduced in the mutant strain EHBR (Mikami et al., 1986), which have a hereditary defect in the cMOAT (Fernandez et al., 1992; Sathirakul et al., 1994; Takenaka et al., 1995a,b; Takikawa et al., 1991; Yamazaki et al., 1996; Chu et al., 1997a,b), compared with that in normal rats. ATP-dependent uptake of GPFX by CMV from normal rats was observed, whereas almost none was observed by CMV from EHBR (Sasabe et al., in press). These results demonstrated that at least part of the GPFX transport across the bile canalicular membrane is mediated by a primary active transport mechanism, cMOAT. Thus both the hepatic uptake and the biliary excretion of GPFX occur by carrier-mediated active transport. Moreover, we have examined the biliary excretion in vivo and the CMV uptake in vitro for the glucuronide from racemic GPFX and found that a major part of the biliary excretion of its 3-glucuronide, a major metabolite, was also mediated by cMOAT (Sasabe et al., in press).

Recently, the stereospecific pharmacokinetics of a number of drugs have been reported. This arises from the stereoselectivity in the drug-metabolizing enzyme, P-450, and/or from a difference in tissue distribution due to the stereoselectivity of plasma protein binding (Wedlund et al., 1985; Meier et al., 1985; Fujimaki, 1992). In terms of carrier-mediated transport, the transport of hexose and amino acids exhibits great stereoselectivity (Shimada and Hoshi, 1986; Jorgensen et al., 1990). Such a difference between optical isomers was also observed in the uptake of -lactam antibiotics by renal brush-border membrane (Kramer et al., 1992).

GPFX has a chiral center on the piperazine ring (fig. 1). The glucuronides of R(+)-GPFX and S()-GPFX (R-GPFX-Glu and S-GPFX-Glu, respectively) are diastereomers because of the chirality in the glucuronic acid moiety. Therefore, if the transporters can discriminate between the stereoisomers of GPFX and its glucuronide, their pharmacokinetics may exhibit differences between the isomers. There has been no study of the stereospecific pharmacokinetics of NQs, except the glucuronidation of OFLX, which was reported to be stereoselective (Okazaki et al., 1989, 1991). Moreover, no stereospecific biliary excretion has ever been observed. Until now, no information has been published on stereoselective recognition by cMOAT. Thus, in order to examine whether stereoselective pharmacokinetics is present, we determined the plasma concentration and biliary excretion at steady state in rats receiving a constant infusion of each GPFX epimer. Furthermore, to clarify the stereoselectivity in the transport mechanism for the isomers, we performed in vitro studies of the hepatic uptake and biliary excretion, using isolated rat hepatocytes and CMV, respectively.

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Chemical structures of GPFX, its 3-glucuronide and OPC-17203 (internal standard).

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Chemicals. Unlabeled R(+)-GPFX and S()-GPFX were synthesized by Otsuka Pharmaceutical Company (Tokyo, Japan). R(+)-GPFX-Glu and S()-GPFX-Glu were isolated from the bile of rats given a constant infusion of the corresponding epimer of GPFX at a dose of 25 µg/min/kg (70 nmol/min/kg). Bile (1 ml) was applied to a pretreatment column (Bond Elut C18, 3 ml/500 mg, Varian, CA) and then eluted with methanol/water (15:85, v/v) after washing with water. The eluate was injected onto an HPLC column (Tosoh, TSK gel ODS-80TM, 150 × 4.6 mm I.D., Tokyo) to purify the 3-glucuronide. DNP-SG was synthesized from 1-chloro-2,4-dinitrobenzene and radiolabeled glutathione (GSH) in the presence of glutathione S-transferase as described by Kobayashi et al. (1990). [³H]S-GSH (1.62 MBq/µmol, radiochemical purity, 99.9%) and [³H]taurocholic acid (TCA, 128 MBq/µmol, 98.5%) were purchased from New England Nuclear (Boston, MA). ATP, creatine phosphate and creatine phosphokinase were purchased from Sigma Chemical Co. (St. Louis, MO). Collagenase was obtained from Wako Pure Chemical Industries Ltd (Osaka, Japan). All other chemicals were of reagent grade.

In vivo study at steady state. Male Sprague-Dawley (SD) rats (Nihon Ikagaku, Tokyo, Japan) and EHBR (SLC Shizuoka, Japan) weighing approximately 250 to 300 g were used throughout the experiments. Under light ether anesthesia, the femoral artery and vein were cannulated with a polyethylene catheter (PE-50) for blood sampling and for the injection of GPFX, respectively. The bile duct was cannulated with a polyethylene catheter (PE-10) for bile collection. The bladder was also cannulated to collect urine. The rats received constant infusions of the GPFX epimers at a dose of 15 µg/min/kg (41.9 nmol/min/kg) following a bolus i.v. administration of 5 mg/kg (13.9 µmol/kg). Bile and urine were collected in preweighed test tubes at 20- and 30-min intervals, respectively, throughout the experiment. Plasma was prepared by centrifugation of the blood samples (10,000 × g; Microfuge, Beckman, Fullerton, CA). The rats were killed after 160 min, and the entire liver and kidneys were excised immediately. The tissues were weighed and stored at 30°C until required for assay. Portions of liver and kidney were added to nine volumes of 50% methanol (v/v) and homogenized. To the homogenate (50 µl) was added an internal standard (OPC-17203, 100 ng) followed by centrifugation in the tabletop microfuge after dilution with 50% methanol (200 µl). The resulting supernatants (20 µl) were subjected to HPLC to determine the concentration of GPFX and its glucuronide. The internal standard (100 ng) was added to plasma (25 µl) together with 50% methanol (200 µl) to precipitate proteins. Bile and urine (10 µl) were diluted with 25% acetonitrile (400 µl).

Plasma protein binding of stereoisomers of GPFX and GPFX-Glu. Plasma protein binding was determined by ultrafiltration. Initially, the filtration tubes (Centrifree MPS-III, Amicon, Tokyo, Japan), in which blank plasma had been placed, were centrifuged to obtain the filtrate (2000 × g for 30 min). The other MPS-III tube was applied to the filtrate obtained and then centrifuged. This pretreatment with MPS-III reduced the nonspecific adsorption of GPFX to the membrane filter. Plasma samples with GPFX or its glucuronide (5 µM) were placed on MPS-III pretreated in this way, following incubation at 37°C for 3 min. Then the plasma was centrifuged (2000 × g for 10 min) to give the filtrate containing unbound compound. The concentrations in the plasma and filtrate were determined as described, using HPLC.

Hepatocyte preparation. Hepatocytes were isolated from male SD rats by the procedure of Baur et al. (1975). After isolation, the hepatocytes were suspended (2 mg protein/ml) in ice-cold albumin-free Krebs-Henseleit buffer supplemented with 12.5 mM HEPES (pH 7.3). The study was carried out in the presence of sodium ions. Cell viability was routinely checked by the trypan blue [0.4% (w/v)] exclusion test. We used over 90% as a viability criterion for the hepatocyte studies. Protein concentrations were determined by the method described by Bradford (1976), using the Bio-Rad protein assay kit with bovine serum albumin as a standard.

Hepatocyte uptake study of GPFX isomers. Uptake of GPFX isomers (50 µM) was initiated by adding the ligand solution with the epimer (0.5 ml) to the preincubated cell suspension (2 mg protein/ml, 0.5 ml) at 37°C for 5 min. At a designated time, the reaction was terminated by separating the cells from the medium using a centrifugal filtration technique (Schwenk, 1980; Sasabe et al., 1997). The amount of GPFX in medium and cells was determined as described, using HPLC (Sasabe et al., 1997). The time dependence of GPFX uptake was plotted as an uptake value (µl/mg protein) obtained by dividing the amount taken up by its concentration in the medium.

Preparation of CMV. CMV were prepared from male SD rats and EHBR, using the method of Meier et al. (1984), slightly modified. After suspension of vesicles in 50 mM Tris-HCl buffer (pH 7.4) containing 250 mM sucrose, the vesicles were frozen in liquid N₂ and stored at 80°C until required. The transport activity of CMV used in this study was also checked by measuring the ATP-dependent uptake of standard substrates, [³H]TCA (1 µM) and [³H]DNP-SG (1 µM), for a 2-min incubation period at 37°C. The purity of the prepared CMVs was evaluated by determining the activities of Mg⁺⁺ ATPase (Schoner et al., 1967), alkaline phosphatase (Yachi et al., 1989) and -GTPase (-GTP assay kit, Wako Pure Chemical Industries, Osaka, Japan). The ratio of inside-out vesicles and right-side-out vesicles was determined by measuring nucleotide pyrophosphatase activity (Bohme et al., 1994).

Uptake of the stereoisomers of GPFX-Glu by CMV. The uptake of GPFX-Glu was measured by the rapid filtration technique described by Ishikawa et al. (1990). The transport medium contained 10 mM Tris-HCl buffer (pH 7.4), 250 mM sucrose and 10 mM MgCl₂ with or without 5 mM ATP and ATP-regenerating system (10 mM creatine phosphate and 100 µg/ml creatine phosphokinase). The transport medium (final volume 200 µl) was mixed with about 20 µl of vesicle suspension (100 µg of protein) and incubated at 37°C. The uptake reaction was stopped by addition of 1 ml of ice-cold buffer containing 100 mM NaCl, 250 mM sucrose and 10 mM Tris-HCl (pH 7.4) at designated times. This reaction mixture (1 ml) was then filtered through a 0.45-µm HAWP filter (Millipore Corp., Bedford, MA) and washed twice with 5 ml of ice-cold buffer. The filter was added into 50% methanol (700 µl, v/v) and shaken vigorously to extract any glucuronide retained. The other portion of the reaction mixture (50 µl) was diluted with 50% methanol (200 µl) to determine the medium concentration. Both methanol solutions were centrifuged, and then the supernatant (20 µl) was subjected to HPLC. The saturation study for the uptake of the glucuronide was performed at a medium concentration of 2.5, 5, 20, 50, 200 and 1000 µM. The uptake of the glucuronide by CMV was normalized with respect to both the amount of vesicles and the medium concentration of ligand. The initial uptake velocity was assessed from the uptake at 2 min when linear uptake was observed (Sasabe et al., in press).

Inhibitory effects of stereoisomers of GPFX and GPFX-Glu on ATP-dependent uptake of [³H]DNP-SG by CMV. The uptake of [³H]DNP-SG (1 µM) was studied by the same method as for GPFX-Glu except for a change in the volume of the incubation mixture (final volume 20 µl) and the amount of vesicles (10 µg of protein). The uptake reaction was stopped after 2 min. For inhibition of the uptake of [³H]DNP-SG, the reaction mixture was incubated in the presence of GPFX isomers at concentrations of 0.1, 1, 3 and 10 mM or in the presence of their glucuronide isomers at concentrations of 1, 3, 10, 30, 100 µM and 1 mM. This reaction mixture (900 µl) was then filtered by the same method. Radioactivity retained on the filter was determined by liquid scintillation counting.

Estimation of kinetic parameters. The kinetic parameters for the uptake of GPFX-Glu by CMV were estimated from the following equation:

(1)

where V₀ is the initial uptake rate of the drug (pmol/min/mg protein), S is the drug concentration in the medium (µM), K_m is the Michaelis-Menten constant (µM), V_max is the maximum uptake rate (pmol/min/mg protein) and P_dif is the nonspecific uptake clearance (µl/min/mg protein). This equation was fitted to the uptake data sets by an iterative nonlinear least-squares method using a MULTI program (Yamaoka et al., 1981) in order to obtain estimates of the kinetic parameters. The input data were weighted as the reciprocal of the observed values, and the Damping Gauss Newton Method algorithm was used for fitting. The fitted line was converted to the V₀/S vs. V₀ form (Eadie-Hofstee plot).

(2)

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Pharmacokinetics of stereoisomers of GPFX at steady state. Epimers of GPFX were separately dosed by constant infusion to rats. The plasma concentration of the parent reached steady state about 1 hr after the beginning of the infusion in both cases. No difference was observed between C_plasma^ss of the optical isomers (table 1). C_liver^ss and C_kidney^ss were much higher than C_plasma^ss (table 1). There was no difference between the optical isomers for the tissue and plasma concentrations (table 1). Furthermore, the biliary and urinary excretion clearances, CL_bile^plasma, CL_bile^liver, CL_urine^plasma and CL_urine^kidney, showed no statistically significant difference between the two optical isomers of GPFX (table 1, fig. 2).

View larger version (25K): [in this window] [in a new window]   Fig. 2.   a) Biliary excretion rates of R-GPFX and S-GPFX and their glucuronides. (b) Bile flow rates during constant infusion (41.9 nmol/min/kg) of R-GPFX or S-GPFX to normal rats. The biliary excretion rates of R-GPFX () and S-GPFX () and the corresponding glucuronides ( and ) were determined during constant infusion of R-GPFX and S-GPFX, respectively. The bile flow rates after administration of R-GPFX and after that of S-GPFX are shown by dotted and solid lines, respectively. Each point plotted, with its vertical bar, represents the mean ± S.E. of three rats.

Plasma protein binding of stereoisomers of GPFX and GPFX-Glu. No statistically significant difference was observed between the unbound fractions of R-GPFX and S-GPFX in plasma, these being 0.706 ± 0.039 and 0.676 ± 0.038, respectively (table 1). There was also no difference in plasma protein binding between the diastereomers of GPFX-Glu (table 1).

Uptake of optical isomers of GPFX by isolated rat hepatocytes. We found that the uptake of R-GPFX and that of S-GPFX by isolated rat hepatocytes were almost identical, and there was no significant difference between the isomers. The values for the uptake of R-GPFX and for that of S-GPFX were 96 ± 2 and 100 ± 3, 155 ± 4 and 161 ± 8, 175 ± 2 and 178 ± 3, 185 ± 5 and 188 ± 5, 186 ± 3 and 182 ± 3, 178 ± 1 and 181 ± 5 µl/mg protein at 15 and 45 sec and 1, 2, 3 and 5 min of incubation (mean ± S.E. of four determinations in two independent preparations), respectively.

Characterization of CMV. The Mg⁺⁺ ATPase, alkaline phosphatase and -GTPase enrichments, compared with the corresponding activity in liver homogenate, were 35.6 ± 7.7, 63.0 ± 13.5 and 80.6 ± 24.2 times (mean ± S.E., N = 4), respectively, that in normal SD rats. The Mg⁺⁺ ATPase and alkaline phosphatase enrichments in EHBR were 75.1 and 36.6 times (N = 1), respectively. The ratio of inside-out vesicles to the whole vesicle fraction was 33.6 ± 0.5% (mean ± S.E., N = 4) in normal SD rats.

Uptake of GPFX-Glu by CMV prepared from normal rats. The time-dependent uptake of R-GPFX-Glu and S-GPFX-Glu by CMV was investigated at a substrate concentration of 50 µM. In the absence of ATP, the uptake of both glucuronides was below the detection limit of the HPLC method. These uptakes were thought to be approximately less than 5 µl/mg, considering the value of the signal/noise ratio on the HPLC chromatogram. On the other hand, the uptake of glucuronides could be observed, exhibiting overshoot phenomena, in the presence of ATP (fig. 3). At 5 min after the beginning of the incubation, the uptake of R-GPFX-Glu was approximately 2 times greater than that of S-GPFX-Glu, and the difference was statistically significant (fig. 3). Concentration-dependent uptake is shown as an Eadie-Hofstee plot in figure 4. The uptake of both R-GPFX-Glu and S-GPFX-Glu consisted of a saturable and a nonsaturable component. The calculated V_max and K_m values for R-GPFX-Glu were 2.9 times and 1.7 times those for S-GPFX-Glu, respectively (table 2).

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Time-dependent uptake of R-GPFX-Glu and S-GPFX-Glu by CMV prepared from normal rats. The uptake of GPFX-Glu was determined by incubating the vesicles (20 µg) at 37°C for 2, 5, 15, 60 and 120 min after the addition of GPFX-Glu. The mixture (40 µl) contained 50 µM R-GPFX-Glu () or S-GPFX-Glu (), 5 mM ATP, 10 mM creatine phosphate and 100 µg/ml creatine phosphokinase. The uptake was calculated by dividing the amount in vesicles by the GPFX-Glu concentration in the medium. The uptake in the absence of ATP was below the detection limit of the method. Considering the signal/noise ratio on the HPLC chromatogram, the ATP-independent uptake would be below about 5 µl/mg (dotted line). Each point plotted, with its vertical bar, represents the mean ± S.E. of three determinations in one preparation from six animals. * P < .05 (significantly different from the uptake of the corresponding S-GPFX-Glu using Student's t test).

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Eadie-Hofstee plot for R-GPFX-Glu and S-GPFX-Glu uptake by CMV prepared from normal rats. The uptake of GPFX-Glu was determined by incubating the vesicles (100 µg) at 37°C for 2 min after the addition of R-GPFX-Glu () or S-GPFX-Glu (). The mixture (200 µl) contained 5, 20, 50, 200 and 1000 µM GPFX-Glu, 5 mM ATP, 10 mM creatine phosphate and 100 µg/ml creatine phosphokinase. The uptake was calculated by dividing the amount in vesicles by the GPFX-Glu concentration in the medium. Each point plotted, with its vertical bar, represents the mean ± S.E. of three determinations in one preparation from six animals.

Inhibitory effects of stereoisomers of GPFX and GPFX-Glu on [³H]DNP-SG uptake by CMV. The inhibition study of [³H]DNP-SG uptake by CMV was performed to determine the affinity of GPFX isomers and of their glucuronides. All compounds inhibited the ATP-dependent uptake of [³H]DNP-SG in a concentration-dependent manner (fig. 5). The K_i values for R-GPFX-Glu and S-GPFX-Glu differed, being 21.5 ± 4.8 µM and 8.83 ± 1.68 µM, respectively (table 2). These K_i values were almost identical with the corresponding K_m values for their own uptake by CMV (table 2). The K_i values for R-GPFX and S-GPFX were 4.39 ± 3.17 mM and 3.70 ± 2.60 mM, respectively, which shows that there is no difference between the optical isomers (fig. 5). The K_i values for isomers of the parent were more than 200 times larger than those for their glucuronides.

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Effects of stereoisomers of GPFX and their glucuronide on the ATP-dependent uptake of [3H]-DNP-SG by CMV. The uptake of [3H]-DNP-SG was determined by incubating the vesicles (10 µg) at 37°C for 2 min after the addition of [3H]-DNP-SG. The mixture (20 µl) contained 1 µM [3H]-DNP-SG, 5 mM ATP, 10 mM creatine phosphate, 100 µg/ml creatine phosphokinase and R-GPFX-Glu () or S-GPFX-Glu () at 0, 0.001, 0.003, 0.01, 0.03, 0.1 and 1 mM. In the case of R-GPFX () or S-GPFX (), the study was performed at concentrations of 0.01, 1, 3 and 10 mM. The uptake was calculated by dividing the amount in vesicles by the DNP-SG concentration in the medium. Each point plotted, with its vertical bar, represents the mean ± S.E. of six determinations from different two experiments.

Comparison of GPFX-Glu uptake by CMV between normal rats and EHBR. The uptake of GPFX-Glu by CMV prepared from EHBR was compared with that by CMV from normal rats (fig. 6). The uptake clearance of R-GPFX-Glu (93.5 ± 2.9 µl/min/mg protein) by CMV from normal rats was greater than that of S-GPFX-Glu (40.1 ± 1.2 µl/min/mg protein) at a concentration of 10 µM (fig. 6A), whereas in the case of EHBR, the uptake clearances of R-GPFX-Glu and S-GPFX-Glu were 21.5 ± 4.4 µl/min/kg and 19.3 ± 0.9 µl/min/kg, respectively (fig. 6B). These uptake clearances for both R-GPFX-Glu and S-GPFX-Glu were significantly reduced in EHBR compared with normal rats. There was thus a marked stereospecificity in CMV of normal rats and none in EHBR (fig. 6). At 100 µM, the uptake clearances of R-GPFX-Glu and S-GPFX-Glu (17.8 ± 0.3 µl/min/mg protein and 5.4 ± 0.4 µl/min/mg protein, respectively) by CMV from normal rats were smaller than at 10 µM (fig. 6A). The uptake clearances by EHBR CMV were 6.7 ± 1.3 µl/min/mg protein and 4.5 ± 0.2 µl/min/mg protein at 100 µM, respectively, and these values were significantly smaller than the corresponding ones at 10 µM. Therefore, saturable uptake of both glucuronides was observed in CMV prepared from normal rats and EHBR (fig. 6B).

View larger version (22K): [in this window] [in a new window]   Fig. 6.   ATP-dependent uptake of stereoisomers of GPFX-Glu by CMV prepared from normal rats (panel a) and EHBR (panel b). The uptake of GPFX-Glu was determined by incubating the vesicles (100 µg) at 37°C for 2 min after the addition of R-GPFX-Glu (hatched bar) or S-GPFX-Glu (open bar). The mixture (200 µl) contained 10 or 100 µM GPFX-Glu with 5 mM ATP, 10 mM creatine phosphate and 100 µg/ml creatine phosphokinase. The uptake clearance was calculated by dividing the initial uptake rate of GPFX-Glu in vesicles by the concentration in the medium. The uptake in the absence of ATP at 10 µM was below the detection limit of the method (<3 µl/min/mg). The ATP-dependent uptake clearance was calculated by subtracting the uptake clearance in the absence of ATP from that in its presence. Each column, with its vertical bar, represents the mean ± S.E. of three determinations in one preparation from six animals. * P < .05, ** P < .01 (significantly different from the uptake at 10 µM using Student's t test). # P < .05, ## P < .01 (significantly different from the uptake by corresponding normal rat CMV using Student's t test).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

GPFX has a chiral center in its chemical structure and therefore has two optical isomers, R-GPFX and S-GPFX. In this study, no difference was observed between the optical isomers of GPFX for the plasma concentration profiles or urinary and biliary excretion rates (table 1), which suggests that the elimination of parent GPFX from the body is not stereoselective. We believe that pharmacokinetic analysis of each individual epimer is needed to investigate any difference between the pharmacokinetics of isomers, because these enantiomers might affect each other's pharmacokinetics. That is, when transporters and/or enzymes are involved in drug disposition, as in the case of GPFX, such interaction between enantiomers may occur. In order to minimize such effects, we investigated the pharmacokinetics of each enantiomer separately after its administration. We consider that there is a minimal possibility that each isomer inhibited the hepatobiliary transport of the other even when racemic GPFX was administered: The K_m of GPFX for the transport system involved in the uptake into liver cells is approximately 150 µM (Sasabe et al., 1997), which is much higher than the actual plasma concentration observed in the in vivo study (table 1). In addition, both the K_i of S-GPFX and that of R-GPFX for the ATP-dependent transport of DNP-SG via biliary excretion were approximately 4 mM, which is also much higher than the liver concentration of these compounds (tables 1 and 2). Therefore, there is little possibility that stereoselective hepatobiliary transport is present in the case of racemic administration. Concerning the pharmacokinetics of the parent compounds, no stereoselective pharmacokinetics was observed for either GPFX isomer in this study (table 1). Moreover, there was no difference between the isomers in plasma protein binding (table 1).

On the other hand, a difference was observed between the biliary excretion rates of diastereomers of GPFX-Glu (fig. 1). Such a difference in the biliary excretion of metabolites can be caused by the following mechanisms: 1) a difference in hepatic uptake of the parents, 2) a difference in the glucuronidation step in hepatocytes and 3) a difference in biliary excretion of the glucuronide via the bile canalicular membrane. In this study, no difference was observed either in the hepatic uptake for each GPFX optical isomer by isolated rat hepatocytes or in their plasma protein binding (fig. 2; table 1). Therefore, the hepatic uptake of the parent may not be the mechanism that differentiates the biliary excretion of these glucuronides. The sum of the biliary and urinary excretion rates of R-GPFX-Glu (29.0 nmol/min/kg and 2.17 nmol/min/kg, respectively) at steady state was 1.6 times greater than that of S-GPFX-Glu (17.4 nmol/min/kg and 2.00 nmol/min/kg, respectively) during constant infusion of the corresponding epimer (table 1), whereas the parents did not differ in C_liver^ss at steady state. Therefore, the glucuronidation clearance of R-GPFX might be greater than that of S-GPFX. To examine specifically the biliary excretion ability of glucuronide, we assessed the CL_bile^liver of these glucuronides. The bile/liver concentration ratios of R-GPFX-Glu and S-GPFX-Glu were 116 and 63, respectively, which implies that R-GPFX-Glu might be more concentratively excreted into bile than S-GPFX-Glu. The CL_bile^liver of R-GPFX-Glu was 1.9 times greater than that of S-GPFX-Glu (table 1). These results show that the difference between the biliary excretion of the diastereomers stems not only from the difference in glucuronide formation but also from the difference in their transport via the canalicular membrane into the bile. In order to assess the biliary excretion of glucuronides more directly, we investigated the uptake of each diastereomer by CMV (fig. 4). The uptake clearance (V_max/K_m) of R-GPFX-Glu by CMV was 1.7 times that of S-GPFX-Glu (table 2). This difference in vitro was comparable with the difference in CL_bile^liver (1.9 times) in the in vivo study. Thus transport via the bile canalicular membrane is more efficient for R-GPFX-Glu than for S-GPFX-Glu.

To clarify the involvement of the active transport system in the biliary excretion of the glucuronide diastereomers, we performed a study of the uptake of these glucuronides by CMV. Both ATP dependence and saturation were observed in the uptake of both diastereomers, which indicates that their uptake is mediated by a primary active transporter (figs. 3 and 4). The K_m value for the uptake of R-GPFX-Glu (17.3 µM) was greater than that of S-GPFX-Glu (10.1 µM), which indicates that S-GPFX-Glu had a higher affinity for the transporter although CL_bile^liver of S-GPFX-Glu was smaller in vivo (tables 1 and 2). However, the V_max value of R-GPFX-Glu was much greater than that of S-GPFX-Glu, the uptake clearance (V_max/K_m) of R-GPFX-Glu (214 µl/min/mg protein) being ultimately greater than that of S-GPFX-Glu (128 µl/min/mg protein). The ratio R/S for V_max/K_m was approximately 1.7, similar to that for CL_bile^liver (1.9) in the in vivo study. These results suggest that the difference between diastereomers in the biliary excretion of GPFX-Glu might be due to the greater capacity of R-GPFX-Glu transport in spite of its lower affinity for an active transporter.

We have already indicated that the biliary excretion of the glucuronide of the racemic GPFX is mediated by the organic anion transporter cMOAT (Sasabe et al., in press). In the present study, we investigated the uptake of R-GPFX-Glu and S-GPFX-Glu, using CMV prepared from normal rats and EHBR, and found that ATP-dependent uptake of both diastereomers was reduced in CMV from EHBR (fig. 6). In addition, an inhibition study was performed on the uptake of [³H]DNP-SG, a glutathione conjugate and a typical substrate for cMOAT (Kobayashi et al., 1990; Yamazaki et al., 1996; Niinuma et al., 1997). The K_i values for the inhibitory effect of R-GPFX-Glu and S-GPFX-Glu on [³H]DNP-SG uptake were almost identical with the corresponding K_m values for their own uptake (table 2; fig. 5). These results demonstrate that the biliary excretion of both of these glucuronides is mediated by cMOAT, considering that DNP-SG uptake is mediated almost exclusively by cMOAT, which is deficient in EHBR (Niinuma et al., 1997). This first report of stereoselective transport by cMOAT demonstrates that cMOAT can discriminate between the diastereomers of its substrate. However, for a more direct demonstration, we need a further uptake experiment using membrane vesicles obtained from cMOAT-transfected cells.

Several reports have shown that stereoisomers of amino acids and glucose, biological compounds, are recognized in different ways by their transporters. A glucose transporter coupled with sodium in the intestine of the hamster transports D-glucose but not L-glucose (Shimada and Hoshi, 1986). A 40-fold difference was found between the affinity of L-leucine and that of D-leucine (K_m = 0.18 mM and 7.22 mM, respectively) for the amino acid transporter that mediates reuptake (reabsorption) at the luminal membrane of the renal proximal tubule in rabbits (Jorgensen et al., 1990). Carrier-mediated transport of lactic acid across Caco-2 cells is stereoselective and exhibits higher affinity and lower capacity for L-lactic acid, compared with the D-isomer (Ogihara et al., 1996). Such a difference in the affinities of the isomers for the transporter is compatible with the previous report that the inhibition of nicotinic acid transport by L-lactic acid was more remarkable than that by D-isomer in brush-border membrane vesicles isolated from the rat small intestine (Simanjuntak et al., 1990). As for exogenous compounds, transport of -lactam antibiotics, recognized by a dipeptide transporter because its three-dimensional structure is similar to that of the dipeptide, was reported to be stereoselective: the uptake of R-cephalexin by brush-border membrane vesicles from rabbit small intestine was greater than that of S-cephalexin (Kramer et al., 1992); and the cis-isomer of ceftibuten was hardly taken up, whereas the trans-isomer underwent extensive uptake (Yoshikawa et al., 1989). The stereoselective transport may greatly affect the pharmacokinetics of these compounds. As for metabolism, the K_m for hydroxylation of S-mephenytoin by human microsomes was approximately 20 µM, whereas that of R-mephenytoin was approximately 400 µM (Meier et al., 1985). The pharmacokinetics of Ofloxacin, an NQ, was also reported to exhibit stereoselectivity because of a difference in the glucuronidation of the optical isomers in hepatocytes (Okazaki et al., 1989; 1991). Although in the present study, a difference was observed in the uptake of diastereomers of GPFX-Glu by CMV, this stereoselectivity was less than that for the transporters mentioned above (table 2). cMOAT is known to transport the glucuronides of many compounds (Elferink et al., 1995; Keppler and Arias, 1997; Kobayashi et al., 1991; Takenaka et al., 1995a,b; Yamazaki et al., 1996; Chu et al., 1997a), so it is reasonable to speculate that the glucuronic acid moiety should be recognized by cMOAT. The chiral center of GPFX is distant from the glucuronic acid moiety (fig. 1). This may be one of the reasons why there is only a small difference between the two GPFX-Glu diastereomers in their transport by cMOAT. Moreover, the V_max values of the diastereomers were different (table 2). Theoretically, V_max is a product of the transporter density and the kinetic constant for the translocation of the complex between a substrate and a transporter. Because the two diastereomers of GPFX-Glu are transported by the same transporter, the numbers of transporters might be equal. Therefore, the kinetic constant for the translocation might be different for each diastereomer.

cMOAT has been reported to mediate the biliary excretion of many types of organic anions, including glucuronide and glutathione conjugates (Ishikawa et al., 1990; Takenaka et al., 1995a; Chu et al., 1997a; Keppler and Arias, 1997). Recently, multiplicity has been clarified in the transporter of such organic anions. The uptake of the glucuronide of E-3040, a dual inhibitor of 5-lipoxygenase and thromboxane A₂ synthetase, by CMV is mediated not only by cMOAT but also by a transporter that is expressed in EHBR (Takenaka et al., 1995a; Niinuma et al., 1997). CPT-11, an anticancer drug, and the glucuronide of its active metabolite, SN-38, were also reported to be excreted by two types of transporters (Chu et al., 1997b). The ATP-dependent uptake of both diastereomers of GPFX-Glu by CMV from EHBR at a concentration of 10 µM was reduced markedly compared with that from normal rats (fig. 6). This result suggests that cMOAT is a major transporter for both diastereomers of GPFX-Glu. However, the ATP-dependence was significant, though small, in the uptake of both glucuronides by CMV from EHBR (fig. 6). In addition, the ATP-dependent uptake of both glucuronides at a concentration of 100 µM in EHBR was smaller than that at 10 µM (fig. 6), which indicates the existence of a saturable transport mechanism in EHBR, too. These results suggest that there may be another transporter in EHBR that mediates the transport of the glucuronides into CMV. No significant difference was observed between the uptakes of two diastereomers of GPFX-Glu by CMV from EHBR (fig. 6), which suggests that this transporter expressed in EHBR cannot discriminate these diastereomers.

We should also note that the concentration dependence in the EHBR CMV study may come from a general toxicity of GPFX-Glu, including utilization of ATP. However, it has also been reported that there is an ATP-dependent and saturable uptake of the glucuronide conjugate of E3040 and SN-38 by CMV from EHBR (Takenaka et al., 1995a; Niinuma et al., 1997; Chu et al., 1997b). On the other hand, most recently, we have found that the cDNA fragment, which was amplified using primers based on the conserved sequence of the carboxy-terminal ATP binding cassette region of human multidrug resistance-associated protein by polymerase chain reaction from SD rat liver, hybridized to poly (A)⁺-RNA from EHBR liver. Fortunately, we have succeeded in cloning cMOAT-like proteins from EHBR liver (Hirohashi et al., 1996). Taking these results into consideration, we believe that a primary active transporter present in EHBR recognizes the glucuronides of several compounds and may also be involved in the GPFX-Glu uptake by EHBR CMV. Further investigation using the expression systems of such a transporter is required for a conclusive demonstration of the primary active transport system of GPFX-Glu in EHBR.

Our present data (table 1) indicate that most of the elimination of GPFX is by glucuronidation, because the total excretion rate (biliary and urinary) is equivalent to about 70% of the infusion rate for R-GPFX and 50% for S-GPFX; the renal excretion of the parent accounts for only 10% of the infusion rate (table 1). Although the total clearances of the two stereoisomers of GPFX are basically identical, the formation of R-GPFX-Glu is greater than that of S-GPFX-Glu. This implies that the residual elimination may also be stereoselective. Such residual elimination includes cleavage of the piperadine ring and sulfation (Akiyama et al., 1995). Sulfation of several compounds has been reported to exhibit stereoselectivity (Walle et al., 1993; Pesola and Walle, 1993). These metabolic pathways for GPFX might also be stereoselective. Further studies have to be performed to demonstrate this hypothesis.

In conclusion, the hepatobiliary transport of the parent GPFX has no stereospecificity, whereas the biliary excretion of R-GPFX-Glu is greater than that of S-GPFX-Glu. This difference between the diastereomers of the parent and glucuronide of GPFX may be due to a difference in the transport activity through the canalicular membrane; R-GPFX-Glu has lower affinity and higher capacity than S-GPFX-Glu. The transporter that mediates such transport is deficient in EHBR. On the other hand, a primary active transport system that is also expressed on the bile canalicular membrane of EHBR exhibited low transport activity for GPFX-Glu and had no stereoselectivity for the diastereomers.